A two-stage approach to localize spatially distributed sources

We present a new technique to localize multiple scattered sources and estimate their angular spreads (ASs). We take into account the partial stationarity of the channel and propose a two-stage approach to estimate the ASs and the nominal angles of arrival (AoAs) of the sources. First, we blindly estimate the channels over several data blocks regularly spaced by intervals larger than the coherence time but each, short enough in length, to make time variations negligible within the block duration. Second, for each spatially scattered source, we separately process the corresponding sequence of quasi-independent channel realization estimates as a new single-scattered-source observation over which we apply Taylor series expansions to transform the estimation of the nominal AoAs and the ASs of the corresponding scattered source into a simple localization of two closely-spaced uncorrelated rays (i.e., point sources). Rayspsila locations allow accurate retrieval of the AoAs and ASs. Simulations confirm the efficiency of the proposed approach in the most adverse conditions.